Switch to pyttsx3 for TTS due to Python 3.13 compatibility issues with Coqui TTS

README.md

@@ -5,7 +5,7 @@ A ComfyUI custom node for text-to-speech integration with a local TTS server.
 ## Features
 
 - Custom ComfyUI node `HttpTTSToAudio` for sending text to a local TTS server.
-- Local TTS server using Coqui TTS (XTTS v2 model).
+- Local TTS server using pyttsx3 (offline TTS engine).
 - Easy setup with batch files for Windows.
 
 ## Prerequisites

install.bat

@@ -3,25 +3,25 @@ echo Setting up TTS environment...
 
 REM Check for Python in ComfyUI root python_embeded directory
 if exist "..\..\..\python_embeded\python.exe" (
-    set PYTHON_EXE="..\..\..\python_embeded\python.exe"
+    set PYTHON_EXE=..\..\..\python_embeded\python.exe
     goto :python_found
 )
 
 REM Check for Python in ComfyUI python_embeded directory (higher in tree)
 if exist "..\..\python_embeded\python.exe" (
-    set PYTHON_EXE="..\..\python_embeded\python.exe"
+    set PYTHON_EXE=..\..\python_embeded\python.exe
     goto :python_found
 )
 
 REM Check for Python in ComfyUI python_embeded directory
 if exist "..\python_embeded\python.exe" (
-    set PYTHON_EXE="..\python_embeded\python.exe"
+    set PYTHON_EXE=..\python_embeded\python.exe
     goto :python_found
 )
 
 REM Check for Python in ComfyUI directory
 if exist "..\python.exe" (
-    set PYTHON_EXE="..\python.exe"
+    set PYTHON_EXE=..\python.exe
     goto :python_found
 )
 
@@ -42,13 +42,13 @@ echo Using Python: %PYTHON_EXE%
 
 REM Set PIP_EXE based on Python location
 if "%PYTHON_EXE%" == "..\..\..\python_embeded\python.exe" (
-    set PIP_EXE="..\..\..\python_embeded\Scripts\pip.exe"
+    set PIP_EXE=..\..\..\python_embeded\Scripts\pip.exe
     set USE_VENV=false
 ) else if "%PYTHON_EXE%" == "..\..\python_embeded\python.exe" (
-    set PIP_EXE="..\..\python_embeded\Scripts\pip.exe"
+    set PIP_EXE=..\..\python_embeded\Scripts\pip.exe
     set USE_VENV=false
 ) else if "%PYTHON_EXE%" == "..\python_embeded\python.exe" (
-    set PIP_EXE="..\python_embeded\Scripts\pip.exe"
+    set PIP_EXE=..\python_embeded\Scripts\pip.exe
     set USE_VENV=false
 ) else (
     set PIP_EXE=pip

requirements.txt

@@ -1,2 +1,2 @@
-TTS
+pyttsx3
 flask

run_tts.bat

@@ -7,13 +7,17 @@ if exist venv (
 ) else (
     REM Check for embedded Python
     if exist "..\..\..\python_embeded\python.exe" (
-        set PYTHON_CMD="..\..\..\python_embeded\python.exe"
-    ) else if exist "..\..\python_embeded\python.exe" (
-        set PYTHON_CMD="..\..\python_embeded\python.exe"
-    ) else if exist "..\python_embeded\python.exe" (
-        set PYTHON_CMD="..\python_embeded\python.exe"
+        set PYTHON_CMD=..\..\..\python_embeded\python.exe
     ) else (
-        set PYTHON_CMD=python
+        if exist "..\..\python_embeded\python.exe" (
+            set PYTHON_CMD=..\..\python_embeded\python.exe
+        ) else (
+            if exist "..\python_embeded\python.exe" (
+                set PYTHON_CMD=..\python_embeded\python.exe
+            ) else (
+                set PYTHON_CMD=python
+            )
+        )
     )
 )
 

tts_server.py

@@ -1,13 +1,10 @@
 from flask import Flask, request, send_file, jsonify
-from TTS.api import TTS
+import pyttsx3
 import os
 import uuid
 
-# Choose a model (example: XTTS v2 multilingual; you can pick any supported model)
-MODEL_NAME = "tts_models/multilingual/multi-dataset/xtts_v2"
-
 app = Flask(__name__)
-tts = TTS(MODEL_NAME)
+engine = pyttsx3.init()
 
 OUTPUT_DIR = "outputs"
 os.makedirs(OUTPUT_DIR, exist_ok=True)
@@ -16,8 +13,7 @@ os.makedirs(OUTPUT_DIR, exist_ok=True)
 def tts_route():
     data = request.get_json()
     text = data.get("text", "").strip()
-    speaker = data.get("speaker")  # optional
-    language = data.get("language", "en")  # optional
+    language = data.get("language", "en")  # pyttsx3 doesn't support language directly, but we can set voice
 
     if not text:
         return jsonify({"error": "No text"}), 400
@@ -25,13 +21,12 @@ def tts_route():
     out_name = f"{uuid.uuid4().hex}.wav"
     out_path = os.path.join(OUTPUT_DIR, out_name)
 
+    # Set properties if needed
+    # engine.setProperty('voice', voice_id)  # for language, but pyttsx3 has limited support
+
     # Generate audio
-    # For XTTS you can pass language, speaker_wav, etc.; adjust to your chosen model
-    tts.tts_to_file(
-        text=text,
-        file_path=out_path,
-        language=language
-    )
+    engine.save_to_file(text, out_path)
+    engine.runAndWait()
 
     return jsonify({
         "ok": True,